I think some of the confusion is coming from the definition of a “broadcast domain”.
A broadcast domain functions on Layer 2 of the OSI model, and is independent of IP addresses and subnets. If you were to send out a packet with the broadcast MAC address (FF:FF:FF:FF:FF:FF), everything on the same broadcast domain sees that packet, and has to check if it is relevant or not. This becomes a big deal with embedded systems since they typically have limited processing power. Packets like ARP are processed even if they are for PC’s on a different subnet. If you are in a broadcast domain with multiple subnets, you can open a network sniffer (like Wireshark) and see broadcast packets from the different subnets.
IP Video can make this even worse if it uses multicast. If you aren’t using switches with IGMP snooping or some other method to isolate traffic (like VLANs), then a couple IP cameras can easily overwhelm small embedded devices with a flood of packets. If if the device isn’t “using” the video, it still receives the packet and has to do some small amount of processing before it knows it can ignore the packet.
VLANs provide us a way to break up broadcast domains. They are very effective, but require managed switches throughout your network. In a VLAN enabled network, there are two types of packets, tagged and untagged. Untagged packets are your typical, every-day ethernet frames. Tagged packets have an 802.1Q VLAN header on them, marking what VLAN they belong to. Likewise, every port in a managed switch has one untagged VLAN assigned to it and potentially one or more tagged VLANS.
Let’s use an exampe from the wiki, here we have two managed switchs linked together and a handful of clients on them.

When the first client sends an untagged packet to the switch, the switch checks its config and sees the port is assigned VLAN1, so it puts a VLAN 1 tag on packet and goes to forward it on. The switch sees no other ports marked as “Untagged VLAN 1”, but it does see an uplink with “Tagged VLAN 1” on it, so it leaves the tag on, and sends the packet out that port. At this point the seconds switch gets the packet and looks at the header. It sees the VLAN 1 tag, and checks its config. It has a port marked “Untagged VLAN 1”, so it strips the VLAN 1 tag off and sends it out the appropriate port. Computers attached to VLAN 2 and VLAN 3 never saw the packet.
VLANs are designed to be transparent to the end device, because the client should only send and receive untagged packets and let the switch handle all the VLAN work. Some care must be taken to only send tagged packets to VLAN aware devices, otherwise some unexpected issues can result. When you send tagged packets to an unmanaged switch, any number of things can happen. The packets can be switched out all ports with the VLAN tag intact, the packet can be discarded as invalid, or the vlan tag can be stripped off and the packet forwarded on as untagged. This last scenario is unfortunately fairly common, and results in your VLANs all being bridged together, thus removing the isolation we were working for.
Unfortunately, with your current design, unless you can upgrade the Netgear JGS524 to a smart switch of some sort, VLANs won’t be very practical or effective for you. I hope that at least this will give you a little bit better clarity as to how VLANs work and fit in to a network design. If it is still a little confusing, don’t feel bad. VLANs are a very hard subject that still trips up a lot of seasoned networking vets (the fact that no two vendors seem to use the same terminology doesn’t really help, either). If you have any other questions, just ask.